xref: /openbmc/u-boot/fs/ubifs/orphan.c (revision b1c0c736)
1 /*
2  * This file is part of UBIFS.
3  *
4  * Copyright (C) 2006-2008 Nokia Corporation.
5  *
6  * This program is free software; you can redistribute it and/or modify it
7  * under the terms of the GNU General Public License version 2 as published by
8  * the Free Software Foundation.
9  *
10  * This program is distributed in the hope that it will be useful, but WITHOUT
11  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
12  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
13  * more details.
14  *
15  * You should have received a copy of the GNU General Public License along with
16  * this program; if not, write to the Free Software Foundation, Inc., 51
17  * Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
18  *
19  * Author: Adrian Hunter
20  */
21 
22 #include "ubifs.h"
23 
24 /*
25  * An orphan is an inode number whose inode node has been committed to the index
26  * with a link count of zero. That happens when an open file is deleted
27  * (unlinked) and then a commit is run. In the normal course of events the inode
28  * would be deleted when the file is closed. However in the case of an unclean
29  * unmount, orphans need to be accounted for. After an unclean unmount, the
30  * orphans' inodes must be deleted which means either scanning the entire index
31  * looking for them, or keeping a list on flash somewhere. This unit implements
32  * the latter approach.
33  *
34  * The orphan area is a fixed number of LEBs situated between the LPT area and
35  * the main area. The number of orphan area LEBs is specified when the file
36  * system is created. The minimum number is 1. The size of the orphan area
37  * should be so that it can hold the maximum number of orphans that are expected
38  * to ever exist at one time.
39  *
40  * The number of orphans that can fit in a LEB is:
41  *
42  *         (c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64)
43  *
44  * For example: a 15872 byte LEB can fit 1980 orphans so 1 LEB may be enough.
45  *
46  * Orphans are accumulated in a rb-tree. When an inode's link count drops to
47  * zero, the inode number is added to the rb-tree. It is removed from the tree
48  * when the inode is deleted.  Any new orphans that are in the orphan tree when
49  * the commit is run, are written to the orphan area in 1 or more orphan nodes.
50  * If the orphan area is full, it is consolidated to make space.  There is
51  * always enough space because validation prevents the user from creating more
52  * than the maximum number of orphans allowed.
53  */
54 
55 /**
56  * tot_avail_orphs - calculate total space.
57  * @c: UBIFS file-system description object
58  *
59  * This function returns the number of orphans that can be written in half
60  * the total space. That leaves half the space for adding new orphans.
61  */
62 static int tot_avail_orphs(struct ubifs_info *c)
63 {
64 	int avail_lebs, avail;
65 
66 	avail_lebs = c->orph_lebs;
67 	avail = avail_lebs *
68 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
69 	return avail / 2;
70 }
71 
72 /**
73  * ubifs_clear_orphans - erase all LEBs used for orphans.
74  * @c: UBIFS file-system description object
75  *
76  * If recovery is not required, then the orphans from the previous session
77  * are not needed. This function locates the LEBs used to record
78  * orphans, and un-maps them.
79  */
80 int ubifs_clear_orphans(struct ubifs_info *c)
81 {
82 	int lnum, err;
83 
84 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
85 		err = ubifs_leb_unmap(c, lnum);
86 		if (err)
87 			return err;
88 	}
89 	c->ohead_lnum = c->orph_first;
90 	c->ohead_offs = 0;
91 	return 0;
92 }
93 
94 /**
95  * insert_dead_orphan - insert an orphan.
96  * @c: UBIFS file-system description object
97  * @inum: orphan inode number
98  *
99  * This function is a helper to the 'do_kill_orphans()' function. The orphan
100  * must be kept until the next commit, so it is added to the rb-tree and the
101  * deletion list.
102  */
103 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
104 {
105 	struct ubifs_orphan *orphan, *o;
106 	struct rb_node **p, *parent = NULL;
107 
108 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
109 	if (!orphan)
110 		return -ENOMEM;
111 	orphan->inum = inum;
112 
113 	p = &c->orph_tree.rb_node;
114 	while (*p) {
115 		parent = *p;
116 		o = rb_entry(parent, struct ubifs_orphan, rb);
117 		if (inum < o->inum)
118 			p = &(*p)->rb_left;
119 		else if (inum > o->inum)
120 			p = &(*p)->rb_right;
121 		else {
122 			/* Already added - no problem */
123 			kfree(orphan);
124 			return 0;
125 		}
126 	}
127 	c->tot_orphans += 1;
128 	rb_link_node(&orphan->rb, parent, p);
129 	rb_insert_color(&orphan->rb, &c->orph_tree);
130 	list_add_tail(&orphan->list, &c->orph_list);
131 	orphan->dnext = c->orph_dnext;
132 	c->orph_dnext = orphan;
133 	dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
134 		c->new_orphans, c->tot_orphans);
135 	return 0;
136 }
137 
138 /**
139  * do_kill_orphans - remove orphan inodes from the index.
140  * @c: UBIFS file-system description object
141  * @sleb: scanned LEB
142  * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
143  * @outofdate: whether the LEB is out of date is returned here
144  * @last_flagged: whether the end orphan node is encountered
145  *
146  * This function is a helper to the 'kill_orphans()' function. It goes through
147  * every orphan node in a LEB and for every inode number recorded, removes
148  * all keys for that inode from the TNC.
149  */
150 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
151 			   unsigned long long *last_cmt_no, int *outofdate,
152 			   int *last_flagged)
153 {
154 	struct ubifs_scan_node *snod;
155 	struct ubifs_orph_node *orph;
156 	unsigned long long cmt_no;
157 	ino_t inum;
158 	int i, n, err, first = 1;
159 
160 	list_for_each_entry(snod, &sleb->nodes, list) {
161 		if (snod->type != UBIFS_ORPH_NODE) {
162 			ubifs_err("invalid node type %d in orphan area at "
163 				  "%d:%d", snod->type, sleb->lnum, snod->offs);
164 			dbg_dump_node(c, snod->node);
165 			return -EINVAL;
166 		}
167 
168 		orph = snod->node;
169 
170 		/* Check commit number */
171 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
172 		/*
173 		 * The commit number on the master node may be less, because
174 		 * of a failed commit. If there are several failed commits in a
175 		 * row, the commit number written on orphan nodes will continue
176 		 * to increase (because the commit number is adjusted here) even
177 		 * though the commit number on the master node stays the same
178 		 * because the master node has not been re-written.
179 		 */
180 		if (cmt_no > c->cmt_no)
181 			c->cmt_no = cmt_no;
182 		if (cmt_no < *last_cmt_no && *last_flagged) {
183 			/*
184 			 * The last orphan node had a higher commit number and
185 			 * was flagged as the last written for that commit
186 			 * number. That makes this orphan node, out of date.
187 			 */
188 			if (!first) {
189 				ubifs_err("out of order commit number %llu in "
190 					  "orphan node at %d:%d",
191 					  cmt_no, sleb->lnum, snod->offs);
192 				dbg_dump_node(c, snod->node);
193 				return -EINVAL;
194 			}
195 			dbg_rcvry("out of date LEB %d", sleb->lnum);
196 			*outofdate = 1;
197 			return 0;
198 		}
199 
200 		if (first)
201 			first = 0;
202 
203 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
204 		for (i = 0; i < n; i++) {
205 			inum = le64_to_cpu(orph->inos[i]);
206 			dbg_rcvry("deleting orphaned inode %lu",
207 				  (unsigned long)inum);
208 			err = ubifs_tnc_remove_ino(c, inum);
209 			if (err)
210 				return err;
211 			err = insert_dead_orphan(c, inum);
212 			if (err)
213 				return err;
214 		}
215 
216 		*last_cmt_no = cmt_no;
217 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
218 			dbg_rcvry("last orph node for commit %llu at %d:%d",
219 				  cmt_no, sleb->lnum, snod->offs);
220 			*last_flagged = 1;
221 		} else
222 			*last_flagged = 0;
223 	}
224 
225 	return 0;
226 }
227 
228 /**
229  * kill_orphans - remove all orphan inodes from the index.
230  * @c: UBIFS file-system description object
231  *
232  * If recovery is required, then orphan inodes recorded during the previous
233  * session (which ended with an unclean unmount) must be deleted from the index.
234  * This is done by updating the TNC, but since the index is not updated until
235  * the next commit, the LEBs where the orphan information is recorded are not
236  * erased until the next commit.
237  */
238 static int kill_orphans(struct ubifs_info *c)
239 {
240 	unsigned long long last_cmt_no = 0;
241 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
242 
243 	c->ohead_lnum = c->orph_first;
244 	c->ohead_offs = 0;
245 	/* Check no-orphans flag and skip this if no orphans */
246 	if (c->no_orphs) {
247 		dbg_rcvry("no orphans");
248 		return 0;
249 	}
250 	/*
251 	 * Orph nodes always start at c->orph_first and are written to each
252 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
253 	 * but may contain out of date orphan nodes if the unmap didn't go
254 	 * through. In addition, the last orphan node written for each commit is
255 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
256 	 * there are orphan nodes from the next commit (i.e. the commit did not
257 	 * complete successfully). In that case, no orphans will have been lost
258 	 * due to the way that orphans are written, and any orphans added will
259 	 * be valid orphans anyway and so can be deleted.
260 	 */
261 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
262 		struct ubifs_scan_leb *sleb;
263 
264 		dbg_rcvry("LEB %d", lnum);
265 		sleb = ubifs_scan(c, lnum, 0, c->sbuf);
266 		if (IS_ERR(sleb)) {
267 			sleb = ubifs_recover_leb(c, lnum, 0, c->sbuf, 0);
268 			if (IS_ERR(sleb)) {
269 				err = PTR_ERR(sleb);
270 				break;
271 			}
272 		}
273 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
274 				      &last_flagged);
275 		if (err || outofdate) {
276 			ubifs_scan_destroy(sleb);
277 			break;
278 		}
279 		if (sleb->endpt) {
280 			c->ohead_lnum = lnum;
281 			c->ohead_offs = sleb->endpt;
282 		}
283 		ubifs_scan_destroy(sleb);
284 	}
285 	return err;
286 }
287 
288 /**
289  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
290  * @c: UBIFS file-system description object
291  * @unclean: indicates recovery from unclean unmount
292  * @read_only: indicates read only mount
293  *
294  * This function is called when mounting to erase orphans from the previous
295  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
296  * orphans are deleted.
297  */
298 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
299 {
300 	int err = 0;
301 
302 	c->max_orphans = tot_avail_orphs(c);
303 
304 	if (!read_only) {
305 		c->orph_buf = vmalloc(c->leb_size);
306 		if (!c->orph_buf)
307 			return -ENOMEM;
308 	}
309 
310 	if (unclean)
311 		err = kill_orphans(c);
312 	else if (!read_only)
313 		err = ubifs_clear_orphans(c);
314 
315 	return err;
316 }
317